other financial flows, such as portfolio investment.3 FDI flows are determined to a large extent by long-term economic fundamentals such as an effective infrastructure, a skilled workforce, macroeconomic stability and political and legal predictability (see for example Wheeler and Mody, 1992), while portfolio investment often responds to short-run investment opportunities. The long-run nature of FDI means that it is less likely to be subject to idiosyncratic shocks than portfolio investment. Such shocks that can lead to abrupt reversals or sudden stops of short-run capital inflows can lower the productivity of existing capital stock, resulting in unexpected swings in relative prices and even lead to banking crisis or corporate bankruptcies, especially in small economies (Milesi-Ferrett and Razin, 1998). A further reason for considering FDI rather than portfolio investment concerns the importance of investment funds for portfolio investment that are to a large extent managed by multilateral financial institutions. It is not clear therefore to what extent such flows can be considered for bilateral risk sharing. For example, a significant share of portfolio investment by US investment funds in a particular region is likely to be from investors in third countries, and is thus not relevant for bilateral risk sharing. The second variable of interest to us is a proxy for employment protection. As discussed above this variable is included since stronger labour market regulation may help enforce the insurance aspect of labour contracts shifting risk from wages to profits, and may make future income more secure, thus easing borrowing restrictions, both of which may increase consumption correlations. This part of our empirical analysis represents an extension of the current literature, which has paid little attention to this potential 3
Imbs (2006) uses the IMF’s Coordinated Portfolio Investment Survey (CPIS) data, but notes that the use of survey-based data and the lack of data on Foreign Direct Investment could create biases.
10